1. |
- Andersson, Jennifer, et al.
(författare)
-
Reinforcement Learning Control of a Forestry Crane Manipulator
- 2021
-
Ingår i: 2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2021). - Prague : IEEE Robotics and Automation Society. - 9781665417150 - 9781665417143 ; , s. 2121-2126
-
Konferensbidrag (refereegranskat)abstract
- Forestry machines are heavy vehicles performing complex manipulation tasks in unstructured production forest environments. Together with the complex dynamics of the on-board hydraulically actuated cranes, the rough forest terrains have posed a particular challenge in forestry automation. In this study, the feasibility of applying reinforcement learning control to forestry crane manipulators is investigated in a simulated environment. Our results show that it is possible to learn successful actuator-space control policies for energy efficient log grasping by invoking a simple curriculum in a deep reinforcement learning setup. Given the pose of the selected logs, our best control policy reaches a grasping success rate of 97%. Including an energy-optimization goal in the reward function, the energy consumption is significantly reduced compared to control policies learned without incentive for energy optimization, while the increase in cycle time is marginal. The energy-optimization effects can be observed in the overall smoother motion and acceleration profiles during crane manipulation.
|
|
2. |
|
|
3. |
- Backman, Anders, et al.
(författare)
-
easyADL – Wearable Support System for Independent Life despite Dementia
- 2006
-
Ingår i: ACM CHI 2006 Workshop onDesigning Technology for People with Cognitive Impairments.
-
Konferensbidrag (refereegranskat)abstract
- This position paper outlines the easyADL project, a two-year project investigating the possibility of using wearable technology to assist people suffering the dementia disease in performing Activities of Daily Living (ADL). An introduction to the egocentric interaction modeling framework is provided and the virtual reality based development methodology is discusssed.
|
|
4. |
- Backman, Sofi, et al.
(författare)
-
Continuous Control of an Underground Loader Using Deep Reinforcement Learning
- 2021
-
Ingår i: Machines. - : MDPI. - 2075-1702. ; 9:10
-
Tidskriftsartikel (refereegranskat)abstract
- The reinforcement learning control of an underground loader was investigated in a simulated environment by using a multi-agent deep neural network approach. At the start of each loading cycle, one agent selects the dig position from a depth camera image of a pile of fragmented rock. A second agent is responsible for continuous control of the vehicle, with the goal of filling the bucket at the selected loading point while avoiding collisions, getting stuck, or losing ground traction. This relies on motion and force sensors, as well as on a camera and lidar. Using a soft actor–critic algorithm, the agents learn policies for efficient bucket filling over many subsequent loading cycles, with a clear ability to adapt to the changing environment. The best results—on average, 75% of the max capacity—were obtained when including a penalty for energy usage in the reward.
|
|
5. |
- Bodin, Kenneth, et al.
(författare)
-
Constraint based particle fluids on GPGPU
- 2011
-
Konferensbidrag (övrigt vetenskapligt/konstnärligt)abstract
- We present a fluid simulation method adapted for stream parallelism on general purpose graphics processingunits (GPGPU). In this method the equations of Navier and Stokes are discretized using particles and kernelfunctions as in Smoothed Particle Hydrodynamics (SPH), but rather than using penalty methods or solving for a divergence free velocity field, incompressibility is enforced using holonomic kinematic constraints [1]. We useone constraint for each smoothed particle stating that the local density should be kept constant. Other constraintsare used for boundary conditions and multiphysics coupling. We also present a viscosity model in which theshear rate at each pseudo particle is constrained to satisfy a given constitutive law. The computation of theconstraint forces, namely, the pressure and the stresses, requires the solution system of linear equations whichhave a sparse, saddle point structure. These are solved using the Uzawa method of preconditioned conjugate gradients (CG) applied directly to the symmetric indefinite matrix. The overall simulation method has its rootsin a discrete variational principle and the SPOOK time stepping scheme for constrained mechanical systems [2].The SPOOK method is second order accurate on the positions and constraints violations, and is stable at largetime-steps, thus often allowing several orders of magnitude larger timesteps in our method compared to intraditional SPH methods. The numerical implementation on GPGPU that is the main result of this paper consistsof the following components: particle neighbour searches based on spatial decomposition; summation of kernel densities; construction of Jacobians representing the constraints on the density, boundary conditions, viscosityand multiphysics couplings; a Uzawa CG solver for the system of linear equations; and finally, discrete timestepping of velocities and positions. The CG solver is particularly suitable for stream computing since it is basedon matrix-vector multiplications. The sparse system data is stored in a compressed matrix format and the algorithms operating on this data on GPGPU are implemented in CUDA and OpenCL. Our simulation resultsinclude performance measurements, and validation of the method for benchmark problems. We achieve up totwo orders of magnitude speed-up from the GPGPU over traditional processors and together with the increased timestep efficiency of our method we arrive at interactive performance for systems with up to two million fluidparticles representing an incompressible fluid.
|
|
6. |
- Bodin, Kenneth, et al.
(författare)
-
Constraint Fluids
- 2012
-
Ingår i: IEEE Transactions on Visualization and Computer Graphics. - Los Alamitos, USA : IEEE Computer Society. - 1077-2626 .- 1941-0506. ; 18:3, s. 516-526
-
Tidskriftsartikel (refereegranskat)abstract
- We present a fluid simulation method whereincompressibility is enforced through a holonomic constrainton the mass density. The method starts in aLagrangian particle formulation where the mass densityand other field quantities are represented by SmoothedParticle Hydrodynamics (SPH) kernel approximations.The density constraint is formulated as a regularizedmanybody constraint and is equivalent to very highsound speed. The system is integrated using a variationaldiscrete-time scheme, SPOOK, that includesconstraint regularization and stabilization. This constraintformulation of SPH enables systematic multiphysicsintegration, between rigid multibody physicsand fluids, where buoyancy falls out naturally. The fluidmodel results in a linear system of equations, whilemore general multiphysics systems result in a mixedlinear complementarity problem (MLCP) and we solvethese using iterative methods. The results demonstratenear perfect incompressibility, vastly improved stability,allowing for large time steps, and two orders of magnitudeimproved computational performance. Proof ofconcept is given for computer graphics applications andinteractive simulations.
|
|
7. |
- Bodin, Kenneth, et al.
(författare)
-
Method for simulating dynamic incompressible fluids using particle based spatial discretization and mass density constraints
- 2009
-
Patent (populärvet., debatt m.m.)abstract
- Disclosed is a method of simulating dynamic fluids comprising a set of pseudo particles usedto represent physical quantities of the fluid, a fluid density computed from a weightedaverage over the pseudo particle, and a density constraint on the fluid density constraining itto be incompressible with the reference density of a real physical fluid that is modeled. Theconstraint stabilization forces of the density constraint govern the dynamics of the pseudoparticles and the simulated fluid. The method comprises a combined time stepping andconstraint stabilization and relaxation method that provides global conservation of physicalsymmetries and therefore robust physical and numerical stability under large constraintviolation, and substantial improvement in efficiency over other known methods.
|
|
8. |
- Fordell, Helena, et al.
(författare)
-
A virtual reality test battery for assessment and screening of spatial neglect
- 2011
-
Ingår i: Acta Neurologica Scandinavica. - : John Wiley & Sons. - 0001-6314 .- 1600-0404. ; 123:3, s. 167-174
-
Tidskriftsartikel (refereegranskat)abstract
- Background – There is a need for improved screening methods for spatial neglect.Aim – To construct a VR-test battery and evaluate its accuracy and usability in patients with acute stroke.Method – VR-DiSTRO consists of a standard desktop computer, a CRT monitor and eye shutter stereoscopic glasses, a force feedback interface, and software, developed to create an interactive and immersive 3D experience. VR-tests were developed and validated to the conventional Star Cancellation test, Line bisection, Baking Tray Task (BTT), and Visual Extinction test. A construct validation to The Rivermead Behavioral Inattention Test, used as criterion of visuospatial neglect, was made. Usability was assessed according to ISO 9241-11.Results – Thirty-one patients with stroke were included, 9/31 patients had neglect. The sensitivity was 100% and the specificity 82% for the VR-DiSTRO to correctly identify neglect. VR-BTT and VR-Extinction had the highest correlation (r2 = 0.64 and 0.78), as well as high sensitivity and specificity. The kappa values describing the agreement between traditional neglect tests and the corresponding virtual reality test were between 0.47–0.85. Usability was assessed by a questionnaire; 77% reported that the VR-DiSTRO was ‘easy’ to use. Eighty-eight percent reported that they felt ‘focused’, ‘pleased’ or ‘alert’. No patient had adverse symptoms. The test session took 15 min.Conclusions – The VR-DiSTRO quickly and with a high accuracy identified visuospatial neglect in patients with stroke in this construct validation. The usability among elderly patients with stroke was high. This VR-test battery has the potential to become an important screening instrument for neglect and a valuable adjunct to the neuropsychological assessment.
|
|
9. |
- Fordell, Helena, et al.
(författare)
-
RehAtt – scanning training for neglect enhanced by multi-sensory stimulation in Virtual Reality
- 2016
-
Ingår i: Topics in Stroke Rehabilitation. - : Taylor & Francis Group. - 1074-9357 .- 1945-5119. ; 23:3, s. 191-199
-
Tidskriftsartikel (refereegranskat)abstract
- Background: There is a lack of effective treatment for neglect. We have developed a new training method, RehAtt (TM). The objective of this study was to determine whether RehAtt (TM) improves spatial attention in chronic neglect after stroke. Methods: RehAtt (TM) consists of a computer with monitor, 3D glasses, and a force feedback interface (Robotic pen) giving sensory motor activation to the contra-lesional arm. The software combines visual scanning training with multi-sensory stimulation in 3D virtual reality (VR) game environment. Fifteen stroke patients with chronic neglect (duration > 6 month) had repeated baseline evaluations to confirm stability of symptoms. There were no test-retest effects for any of the tests. Thereafter, all patients trained 15 h in RehAtt (TM) (3 x 1 h for 5 weeks). A neglect test battery and Catherine Bergego Scale, CBS, were used to assess behavioral outcome after intervention. CBS was also used at a 6-month follow-up. Results: Using repeated measurement analysis improvements due to the training were found for Star cancellation test (p = 0.006), Baking tray task (p < 0.001), and Extinction test (p = 0.05). In the Posner task improvements were seen fewer missed targets (p = 0.024). CBS showed improvements in activities of daily life immediately after training (p < 0.01). After 6 months the patients still reported improvement in CBS. Conclusion: RehAtt (TM) is a new concept for rehabilitation of neglect. Training with the VR-method improved spatial attention and showed transfer to improved spatial attention in activities of daily living in chronic neglect. Our results are promising and merit further studies.
|
|
10. |
|
|
11. |
- Servin, Martin, 1975-, et al.
(författare)
-
Examining the smooth and nonsmooth discrete element approaches to granular matter
- 2014
-
Ingår i: International Journal for Numerical Methods in Engineering. - : John Wiley & Sons. - 0029-5981 .- 1097-0207. ; 97:12, s. 878-902
-
Tidskriftsartikel (refereegranskat)abstract
- The smooth and nonsmooth approaches to the discrete element method (DEM) are examined from a computational perspective. The main difference can be understood as using explicit versus implicit time integration. A formula is obtained for estimating the computational effort depending on error tolerance, system geometric shape and size, and on the dynamic state. For the nonsmooth DEM (NDEM), a regularized version mapping to the Hertz contact law is presented. This method has the conventional nonsmooth and smooth DEM as special cases depending on size of time step and value of regularization. The use of the projected Gauss-Seidel solver for NDEM simulation is studied on a range of test systems. The following characteristics are found. First, the smooth DEM is computationally more efficient for soft materials, wide and tall systems, and with increasing flow rate. Secondly, the NDEM is more beneficial for stiff materials, shallow systems, static or slow flow, and with increasing error tolerance. Furthermore, it is found that just as pressure saturates with depth in a granular column, due to force arching, also the required number of iterations saturates and become independent of system size. This effect make the projected Gauss-Seidel solver scale much better than previously thought.
|
|
12. |
- Servin, Martin, et al.
(författare)
-
Hybrid, multiresolution wires with massless frictional contacts
- 2011
-
Ingår i: IEEE Transactions on Visualization and Computer Graphics. - Los Alamitos : IEEE Computer Society. - 1077-2626 .- 1941-0506. ; 7:7, s. 970-982
-
Tidskriftsartikel (refereegranskat)abstract
- We describe a method for the visual interactivesimulation of wires contacting with rigid multibodies. Thephysical model used is a hybrid combining lumped elementsand massless quasistatic representations. The latter is basedon a kinematic constraint preserving the total length of thewire along a segmented path which can involve multiple bodiessimultaneously and dry frictional contact nodes used for roping,lassoing and fastening. These nodes provide stick and slidefriction along edges of the contacting geometries. The lumpedelement resolution is adapted dynamically based on local stabilitycriteria, becoming coarser as the tension increases, and up to thepurely kinematic representation. Kinematic segments and contactnodes are added and deleted and propagated based on contactgeometries and dry friction configurations. The method givesdramatic increase on both performance and robustness becauseit quickly decimates superfluous nodes without loosing stability,yet adapts to complex configurations with many contacts andhigh curvature, keeping a fixed, large integration time step.Numerical results demonstrating the performance and stabilityof the adaptive multiresolution scheme are presented along withan array of representative simulation examples illustrating theversatility of the frictional contact model.
|
|
13. |
- Servin, Martin, et al.
(författare)
-
Visual Simulation of Machine Concepts for Forest Biomass Harvesting
- 2008
-
Ingår i: VRIC '08 Laval Virtual. ; , s. 211-215
-
Konferensbidrag (övrigt vetenskapligt/konstnärligt)abstract
- We present work-in-progress of applying simulator technology for developing, demonstrating and evaluating new machine concepts and work methods in off-road environments. In particular we consider a new concept machine for making efficient biomass harvesting in the process of thinning in dense forest stands and we describe how this machine concept is realized in a virtual environment and plan for simulator experiments.
|
|
14. |
- Sjölie, Daniel, et al.
(författare)
-
Effects of interactivity and 3D-motion on mental rotation brain activity in an immersive virtual environment
- 2010
-
Ingår i: Proceedings of the 28th international conference on Human factors in computing systems. - New York, NY, USA : Association for Computing Machinery (ACM). - 9781605589299 ; , s. 869-878
-
Konferensbidrag (refereegranskat)abstract
- The combination of virtual reality (VR) and brain measurements is a promising development of HCI, but the maturation of this paradigm requires more knowledge about how brain activity is influenced by parameters of VR applications. To this end we investigate the influence of two prominent VR parameters, 3d-motion and interactivity, while brain activity is measured for a mental rotation task, using functional MRI (fMRI). A mental rotation network of brain areas is identified, matching previous results. The addition of interactivity increases the activation in core areas of this network, with more profound effects in frontal and preparatory motor areas. The increases from 3d-motion are restricted to primarily visual areas. We relate these effects to emerging theories of cognition and potential applications for brain-computer interfaces (BCIs). Our results demonstrate one way to provoke increased activity in task-relevant areas, making it easier to detect and use for adaptation and development of HCI.
|
|
15. |
- Sjölie, Daniel, et al.
(författare)
-
Using brain imaging to assess interaction in immersive VR
- 2009
-
Ingår i: Challenges in the evaluation of usability and user experience in reality based interaction. - Boston, MA, USA : ACM. - 9782917490099 ; , s. 23-27
-
Konferensbidrag (refereegranskat)abstract
- We have developed a system where the combination of functional brain imaging (fMRI) and Virtual Reality (VR) can be used to study and evaluate user experience based on brain activation and models of cognitive neuroscience. The ability to study the brain during natural interaction with an (ecologically valid) environment has great potential for several areas of research and development, including evaluation of Reality-Based Interaction (RBI). The RBI concept of tradeoffs is of particular interest since we want to further explore the relation between how the brain works with an accepted reality and what happens when this reality is disrupted. We present the system with an overview of conducted studies to illustrate capabilities and feasibility. In particular, feasibility is supported by the fact that the brain activations seen in these studies match expectations based on existing literature. Further discussion elaborates on the relation to RBI and evaluation; and finally some possible future work is presented.
|
|